10. Power Electronics, Control and Energy Management

Introduction and Context

Renewable-energy systems rarely deliver electricity in the form required by a load or utility network. Photovoltaic modules produce variable direct current, wind turbines generate electricity whose voltage and frequency depend on rotor speed, and batteries operate over a changing voltage range. Modern power systems therefore require electronic conversion, measurement, control, protection, and coordinated energy management. These functions determine not only whether renewable energy can be connected, but also how efficiently, reliably, and safely it can be used.

Power electronics provides the hardware for processing electrical energy, while control engineering determines how that hardware behaves. Energy management operates at a higher level by deciding when energy should be generated, stored, curtailed, purchased, or supplied to a load. In a hybrid renewable system, these layers must work together. A converter may regulate a dc-link voltage, a supervisory controller may allocate power between a battery and a diesel generator, and a protection system may disconnect the installation when abnormal current or voltage is detected.

This chapter develops these relationships and applies them to photovoltaic, wind, battery, and hybrid systems. Particular attention is given to maximum-power-point tracking (MPPT), inverter control, generator control, power quality, protection, and optimization-based energy management.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the roles of dc–dc converters, inverters, rectifiers, and bidirectional converters in renewable systems.

2. Select and interpret appropriate MPPT, generator-control, and inverter-control methods.

3. Distinguish inner-loop control from supervisory energy management.

4. Assess power-quality, protection, grounding, and grid-interconnection requirements.

5. Formulate basic optimization models for dispatching renewable sources, storage, and backup generators.

6. Evaluate practical trade-offs involving efficiency, dynamic response, reliability, cost, and battery lifetime.

The Conversion and Control Architecture

A useful architecture consists of four interacting layers. The electrical-power layer contains switches, magnetic components, capacitors, filters, machines, and protection devices. The measurement layer acquires voltage, current, frequency, temperature, irradiance, state of charge (SOC), and other variables. The fast control layer generates switching commands, usually from pulse-width modulation (PWM). The supervisory layer determines operating modes and power references over intervals ranging from seconds to hours.

A photovoltaic array connected to a common dc link may use a boost converter to raise and regulate voltage. A voltage-source inverter then converts dc power into controlled three-phase ac. A battery is connected through a bidirectional buck–boost converter, allowing charging and discharging. At the point of common coupling, filters and synchronization controls ensure that injected current satisfies grid requirements.

The central engineering principle is separation of time scales. Semiconductor switching may occur at several kilohertz, current-control loops respond within milliseconds, voltage-control loops respond more slowly, and economic dispatch may be updated every few minutes. Poor coordination between these layers can cause oscillation, unnecessary battery cycling, or loss of stability even when each individual controller appears satisfactory.

Important Concepts and Terminology

A converter changes electrical voltage, current, frequency, or direction. A dc–dc converter processes direct current at different voltage levels; an inverter converts dc to ac; a rectifier converts ac to dc; and a bidirectional converter permits power flow in both directions. The dc link is an intermediate energy bus whose voltage is maintained within a prescribed range.

Duty ratio is the fraction of a switching period for which a controlled switch conducts. In an ideal boost converter, the output voltage is approximately V_o = V_i/(1-D), where D is the duty ratio. Real converters experience conduction loss, switching loss, magnetic loss, dead-time distortion, and electromagnetic interference.

MPPT is the process of operating a renewable source at the electrical operating point that maximizes instantaneous power. The maximum-power point changes with irradiance, temperature, wind speed, and operating conditions. SOC represents the estimated energy remaining in a battery, while state of health describes its ageing condition and remaining capacity.

The point of common coupling (PCC) is the electrical interface between a distributed generator and the grid. Power factor is the ratio of active power to apparent power. Total harmonic distortion (THD) measures waveform distortion relative to the fundamental component. Islanding is the continued energization of a section of a grid after its utility supply has been disconnected.

Maximum-Power-Point Tracking

For a photovoltaic module, output power is P = VI. The maximum-power point occurs where dP/dV = 0. Since P = VI, the condition can also be written as I + V(dI/dV) = 0. This incremental-conductance relationship provides the basis for one important MPPT method.

The perturb-and-observe method periodically changes the operating voltage and observes the resulting power change. If power increases, the perturbation continues in the same direction; if power decreases, its direction is reversed. The method is simple and inexpensive, but it oscillates around the maximum and may make an incorrect decision during rapidly changing irradiance.

Incremental conductance estimates dI/dV and can distinguish a voltage perturbation from a change caused by irradiance. It generally performs better under rapidly changing conditions, although it requires more measurement and computation. Other approaches include fractional open-circuit voltage, fractional short-circuit current, fuzzy control, neural estimators, and optimization-based methods. Advanced methods can improve transient performance but introduce parameter, sensor, and validation requirements.

In partially shaded photovoltaic arrays, the power–voltage curve may contain several local maxima. Conventional hill-climbing may settle at a local rather than global maximum. Global-search methods, array reconfiguration, or distributed module-level converters may be required. The designer must compare additional cost and switching loss with the expected energy benefit.

Wind-energy conversion also involves a maximum-power condition. Below rated wind speed, variable-speed operation commonly seeks an optimal tip-speed ratio, defined as rotor peripheral speed divided by wind speed. A torque reference proportional to the square of rotor speed is often used when the turbine characteristic is known. Above rated wind speed, pitch control limits aerodynamic power and protects the machine.

Generator and Inverter Control

Permanent-magnet synchronous generators are frequently connected through a machine-side converter and a grid-side converter. The machine-side controller regulates generator torque or speed, thereby extracting available wind power. The grid-side controller regulates dc-link voltage and controls active and reactive current delivered to the network.

Vector control transforms three-phase quantities into rotating d–q coordinates. With suitable alignment, one current component controls active power or torque and the other controls reactive power or flux. Decoupling terms compensate for cross-coupling between axes. Direct torque control is an alternative that provides rapid torque response but may produce variable switching frequency and greater ripple unless improved modulation is used.

Grid-connected inverters commonly use a phase-locked loop to estimate grid angle. A synchronous-reference-frame current controller compares measured d–q currents with references and generates voltage commands. An outer dc-link loop supplies the active-current reference, while a reactive-power or power-factor loop supplies the reactive-current reference. In weak grids, conventional phase-locked loops can interact adversely with grid impedance; grid-forming or improved grid-following controls may then be preferable.

Grid-forming inverters regulate an internal voltage magnitude and frequency and can establish a stable reference in an islanded microgrid. Grid-following inverters inject current according to an externally detected voltage angle and normally require a sufficiently strong voltage source. This distinction is increasingly important as synchronous generation is displaced by converter-interfaced resources.

Energy Storage, Supervisory Control, and Energy Management

Fast converter controls should not be asked to make economic decisions, and an energy-management system should not directly generate switching pulses. A practical hierarchy uses local protection and current loops at the lowest level, source and storage controllers at an intermediate level, and supervisory dispatch at the highest level.

A battery power balance may be expressed as

P_load = P_PV + P_wind + P_gen + P_dis - P_ch - P_grid,

with consistent sign conventions. Battery SOC evolves approximately according to

SOC(k+1) = SOC(k) + [η_ch P_ch(k) - P_dis(k)/η_dis]Δt/E_nom,

where η_ch and η_dis are charge and discharge efficiencies and E_nom is nominal energy capacity. Practical models also impose SOC limits, charge and discharge power limits, ramp constraints, and reserve requirements.

Rule-based energy management is transparent and easy to implement. A typical strategy serves the load from renewable sources first, charges the battery when surplus exists, discharges it during deficits, and starts a generator when SOC is low or the deficit persists. Such rules are suitable for small systems but may produce excessive cycling or inefficient generator operation.

Optimization-based energy management formulates dispatch as a constrained decision problem. A simplified objective over a horizon may be

min Σ [c_grid P_grid + c_fuel P_gen + c_start y_start + c_deg |P_bat| + c_unserved P_unserved],

where costs represent electricity, fuel, generator starts, battery degradation, and unmet load. Constraints include power balance, generation limits, minimum generator loading, ramp rates, SOC bounds, and reserve margins. Linear programming is useful when relationships are linearized; mixed-integer linear programming represents generator commitment and start-up decisions; nonlinear or model-predictive control may better capture converter and battery behaviour.

Model-predictive control repeatedly optimizes over a forecast horizon, implements only the first control action, updates measurements and forecasts, and solves again. It can coordinate renewable forecasts, storage, flexible loads, and grid prices. Its quality depends on forecast accuracy, computational capability, model fidelity, and a safe fallback mode.

Power Quality and Protection

Renewable converters can affect voltage regulation, frequency response, harmonics, flicker, dc injection, and fault current. PWM produces switching harmonics, which are attenuated through L, LC, or LCL filters. LCL filters provide strong attenuation with smaller inductance but require damping to prevent resonance. Active damping can use measured capacitor current or virtual impedance; passive damping is simpler but dissipates energy.

Reactive-power control can support voltage regulation, while active-power curtailment may be necessary when network voltage exceeds limits. Frequency-support functions can modify active power in response to frequency deviations, provided that energy storage or curtailed renewable power is available.

Protection must address overcurrent, short circuit, overvoltage, undervoltage, overfrequency, underfrequency, insulation failure, ground faults, thermal overload, and loss of synchronization. Semiconductor devices may fail within microseconds, so gate blocking, desaturation detection, fuses, and coordinated hardware protection are essential. Software protection is valuable but should not be the sole defence against high-energy faults.

Anti-islanding protection detects abnormal voltage or frequency, active disturbances, impedance changes, or loss of grid conditions. Inverters must also support controlled reconnection after voltage and frequency have remained within permitted limits. Requirements differ by jurisdiction and interconnection category; therefore, standards, utility rules, and certification procedures must be consulted during design rather than after construction.

Practical Design Process

A robust design begins with resource and load assessment. The engineer identifies operating ranges, worst-case irradiance or wind conditions, motor-starting currents, load criticality, and grid characteristics. Converter ratings are then selected for continuous and transient power, voltage stress, thermal environment, overload capability, and fault conditions.

Control objectives should be stated quantitatively: dc-link voltage deviation, current tracking error, settling time, overshoot, THD, power factor, SOC range, and allowable unserved energy. Sensors require suitable bandwidth, isolation, accuracy, and redundancy where failure could create a hazardous state. Digital controllers should include sampling synchronization, anti-windup, saturation handling, fault logging, and deterministic timing.

Validation should progress from analytical models to averaged simulation, switching simulation, controller-hardware-in-the-loop testing, and finally staged field commissioning. Testing must include irradiance steps, wind changes, load transients, grid disturbances, communication loss, sensor bias, battery limits, and emergency shutdown. A controller that works only under nominal conditions is not an engineered controller.

Illustrative Case Study: Hybrid Microgrid for a Remote Research Station

Background

Consider an illustrative remote research station supplied by a 120 kW photovoltaic array, a 60 kW wind turbine, a 200 kWh lithium-ion battery, and a 100 kW diesel generator. The average load is 75 kW, but refrigeration and communication equipment require continuous supply. The station has no utility connection, and fuel delivery is expensive.

Situation and Problem

A preliminary rule-based controller charged the battery whenever renewable power exceeded load and discharged it whenever a deficit occurred. Operators observed frequent generator starts, operation at low generator loading, and battery SOC values near its limits. During cloudy, calm periods, the system occasionally shed noncritical loads.

Analysis

The main problem was not insufficient installed generation but poor coordination. The controller reacted to instantaneous power rather than forecasted renewable availability. It also ignored generator minimum loading and battery degradation. A generator operating inefficiently for short intervals increased fuel use, while deep and frequent battery cycling reduced available storage capacity.

Approach and Solution

A supervisory mixed-integer optimization model was implemented with a six-hour rolling horizon. Its objective included fuel consumption, generator starts, estimated battery degradation, and a high penalty for unserved critical load. Constraints imposed generator minimum loading, ramp limits, SOC limits of 20–90%, a reserve requirement, and maximum battery charge and discharge power. Local converters retained responsibility for current control, dc-link regulation, and protection.

The strategy allowed the generator to remain off when forecast renewable energy and battery reserve were adequate. When a start was necessary, dispatch kept it above its efficient minimum loading and used surplus power to charge the battery. During likely renewable shortages, the controller preserved SOC rather than discharging aggressively. If communication or optimization failed, a deterministic local rule restored safe operation.

Outcome

The principal outcome was improved coordination: fewer generator starts, less low-load operation, and reduced risk of critical-load interruption. Exact numerical savings would depend on actual weather, fuel curves, battery ageing data, and load patterns; therefore, these outcomes should be verified through site measurements rather than assumed from simulation alone.

Learning Points

The case demonstrates that energy management is different from converter control. It also shows why forecasts, operational constraints, degradation cost, reserve requirements, and fallback modes must be included in a professional design. Optimization is valuable only when its model is realistic and its decisions remain subordinate to safety protection.

Challenges, Risks, and Common Mistakes

A frequent mistake is selecting a converter only from nominal power. Transient overload, thermal derating, switching frequency, altitude, enclosure temperature, and fault energy may determine the actual rating. Another mistake is tuning a controller without considering sensor delay, sampling delay, PWM delay, saturation, or filter resonance.

MPPT can become unstable when the converter current limit is reached or when partial shading creates multiple peaks. SOC estimates may drift because of current-sensor bias, temperature effects, and uncertain battery efficiency. Energy-management models may be overconfident when forecasts are poor. Excessively rapid dispatch can cause unnecessary battery cycling, whereas excessively slow dispatch may miss renewable availability.

Communication networks introduce cyber and operational risks. A lost measurement should not be interpreted as zero power, and a corrupted command should not bypass local limits. Systems require authentication, role-based access, secure updates, event records, and manual emergency controls. Protection settings must be coordinated so that faults are isolated selectively without causing unnecessary system-wide shutdown.

Professional and Regulatory Considerations

Engineers must design according to applicable grid-interconnection, electromagnetic-compatibility, battery-safety, machinery, and electrical-installation requirements. Responsibilities include documenting assumptions, verifying protection settings, preserving test records, and communicating limitations to operators. Energy-management algorithms should not conceal curtailment, degraded storage, or unmet-load risk. Where automated decisions affect essential services, human override and clear alarm priorities are necessary.

Current Developments

Wide-bandgap semiconductors such as silicon carbide and gallium nitride can reduce switching loss and increase power density, although cost, gate-drive design, insulation, and electromagnetic compatibility remain important. Digital twins and hardware-in-the-loop platforms are improving controller validation. Forecast-assisted optimization increasingly combines weather prediction, probabilistic reserve estimation, and battery-health models.

Virtual power plants coordinate distributed batteries, electric vehicles, flexible loads, and renewable generators. Artificial intelligence can assist forecasting, anomaly detection, and adaptive control, but data quality, explainability, cyber security, and safe constraint enforcement must be addressed. Grid-forming control, synthetic inertia, and black-start-capable converter systems are becoming important as inverter-based generation increases.

Best-Practice Design Principles

1. Separate protection, fast control, source control, and supervisory optimization.

2. Define operating limits and safe fallback modes before optimizing energy cost.

3. Validate controllers under abnormal as well as nominal conditions.

4. Include battery degradation, reserve, minimum generator loading, and forecast uncertainty in dispatch decisions.

5. Design filters, grounding, shielding, and measurement systems as part of the converter rather than as afterthoughts.

6. Record assumptions, calibration data, protection settings, software versions, and commissioning results.

7. Treat cyber security, maintainability, operator training, and regulatory compliance as engineering requirements.

Chapter Summary

Renewable-energy systems depend on coordinated power conversion and control. Dc–dc converters condition source and storage voltages, inverters regulate active and reactive power, and machine-side controls extract energy from wind generators. MPPT methods seek the instantaneous operating point of maximum renewable power, while hierarchical control separates rapid electrical regulation from slower supervisory decisions.

Energy management balances renewable generation, storage, backup generation, loads, and grid exchange subject to technical and economic constraints. Rule-based methods are simple, whereas mixed-integer optimization and model-predictive control can account for generator commitment, battery degradation, forecasts, and reserve. Power quality, protection, anti-islanding, grounding, thermal limits, and cyber security are essential parts of the system rather than secondary details. Effective design combines accurate models, robust local control, optimization, staged testing, and professionally governed operation.

Key Terms

Converter: Power-electronic equipment that changes electrical voltage, current, frequency, or direction.

Inverter: A converter that changes dc power into ac power.

Dc link: An intermediate dc bus connecting conversion stages.

Duty ratio: The fraction of a switching cycle during which a switch conducts.

MPPT: Control that operates a renewable source near its maximum-power point.

Incremental conductance: An MPPT method using the condition I + V(dI/dV) = 0.

Tip-speed ratio: The ratio of blade-tip speed to wind speed.

Vector control: Control using rotating d–q coordinates to regulate decoupled power variables.

Grid-following inverter: An inverter that synchronizes to an existing grid voltage waveform.

Grid-forming inverter: An inverter that establishes voltage magnitude and frequency.

SOC: Estimated fraction of usable battery energy remaining.

State of health: Measure of battery ageing and remaining performance capability.

PCC: Point where a distributed generator connects to the external grid.

THD: Measure of harmonic distortion relative to the fundamental waveform.

Anti-islanding: Functions that prevent unintended energization of an isolated grid section.

Model-predictive control: Receding-horizon optimization that repeatedly updates control actions.

Reserve margin: Available extra generation or storage capacity held for uncertainty or contingencies.

LCL filter: Converter output filter using two inductors and a capacitor to attenuate switching harmonics.

Review Questions

Short Answer Questions

1. What functions are performed by a bidirectional dc–dc converter in a battery energy-storage system?

2. State the maximum-power condition for a photovoltaic source in incremental-conductance form.

3. Distinguish grid-following and grid-forming inverter control.

4. Why is a reserve margin included in hybrid-system energy management?

5. Identify two hardware protection functions required in a high-power converter.

Long Answer Questions

1. Explain the hierarchical control architecture of a renewable hybrid microgrid and discuss the time-scale separation between its layers.

2. Compare perturb-and-observe and incremental-conductance MPPT, including their behaviour under rapidly changing irradiance.

3. Describe vector control of a wind-energy conversion system with separate machine-side and grid-side converters.

4. Formulate and explain an optimization-based dispatch problem for photovoltaic generation, a battery, and a diesel generator.

5. Discuss the main sources of power-quality disturbance in inverter-based renewable systems and methods for mitigation.

Analytical / Application Questions

1. A photovoltaic converter oscillates around its operating point during stable irradiance but tracks poorly during fast cloud movement. Diagnose the likely limitation and recommend a control improvement.

2. A remote microgrid starts its generator frequently and operates it at low load. Propose revised supervisory constraints and objective terms to improve operation without compromising reserve.

3. An inverter-based microgrid must operate during utility outages. Decide where grid-forming control, anti-islanding protection, battery SOC limits, and load-shedding logic should be coordinated, and justify the sequence of operation.

22. Multiple Choice Questions with Answers

Question 1

For a photovoltaic array, the maximum-power point is characterized by which condition?

A. I − V(dI/dV) = 0

B. I + V(dI/dV) = 0

C. V + I(dV/dI) = 0 for every operating point

D. dV/dI = 0 only

Correct Answer: B. I + V(dI/dV) = 0

Question 2

What is the principal weakness of basic perturb-and-observe MPPT during rapidly changing irradiance?

A. It cannot measure voltage

B. It always requires a wind-speed sensor

C. It may confuse irradiance change with the effect of its perturbation

D. It operates only with batteries

Correct Answer: C. It may confuse irradiance change with the effect of its perturbation

Question 3

In a grid-connected inverter using d–q control, the d-axis current is commonly assigned to control:

A. Active power or dc-link power balance

B. Filter temperature only

C. Ground-fault resistance

D. Switching-device insulation

Correct Answer: A. Active power or dc-link power balance

Question 4

Which function is normally assigned to the grid-side converter of a variable-speed wind turbine?

A. Changing wind direction

B. Regulating dc-link voltage and controlling grid current

C. Measuring blade erosion only

D. Setting the turbine foundation load

Correct Answer: B. Regulating dc-link voltage and controlling grid current

Question 5

Why does an energy-management system impose a minimum loading constraint on a diesel generator?

A. To prevent all renewable generation

B. To ensure the generator operates only at zero speed

C. To avoid inefficient and potentially damaging low-load operation

D. To eliminate the need for a battery

Correct Answer: C. To avoid inefficient and potentially damaging low-load operation

Question 6

Which statement best describes model-predictive control in a hybrid renewable system?

A. It calculates switching pulses once during commissioning

B. It optimizes over a forecast horizon and repeatedly implements the next action

C. It disconnects all loads whenever SOC changes

D. It replaces hardware protection

Correct Answer: B. It optimizes over a forecast horizon and repeatedly implements the next action

Question 7

What is the main purpose of an LCL filter at an inverter output?

A. To increase wind speed

B. To attenuate switching harmonics before grid connection

C. To estimate battery SOC

D. To provide fuel to a generator

Correct Answer: B. To attenuate switching harmonics before grid connection

Question 8

Which protection issue is specifically associated with a distributed inverter continuing to energize a disconnected grid section?

A. Thermal runaway

B. Islanding

C. Tip-speed tracking

D. Duty-ratio saturation

Correct Answer: B. Islanding

Question 9

A supervisory controller should normally respond more slowly than an inner current loop because it:

A. Makes economic and operating-mode decisions rather than regulating switching instantaneously

B. Must always ignore measurements

C. Controls only semiconductor temperature

D. Cannot use forecasts

Correct Answer: A. Makes economic and operating-mode decisions rather than regulating switching instantaneously

Question 10

Which addition most directly improves an optimization model for battery dispatch?

A. Removing SOC limits

B. Ignoring charge efficiency

C. Including degradation cost, SOC bounds, and charge/discharge power limits

D. Allowing unlimited simultaneous charging and discharging without penalty

Correct Answer: C. Including degradation cost, SOC bounds, and charge/discharge power limits